|
{ |
|
"best_metric": null, |
|
"best_model_checkpoint": null, |
|
"epoch": 9.6, |
|
"eval_steps": 500, |
|
"global_step": 120, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 0.4, |
|
"grad_norm": 0.46419957280158997, |
|
"learning_rate": 0.0001999633286223284, |
|
"loss": 0.5697, |
|
"step": 5 |
|
}, |
|
{ |
|
"epoch": 0.8, |
|
"grad_norm": 0.47279733419418335, |
|
"learning_rate": 0.00019868265225415265, |
|
"loss": 0.581, |
|
"step": 10 |
|
}, |
|
{ |
|
"epoch": 1.2, |
|
"grad_norm": 0.34838807582855225, |
|
"learning_rate": 0.00019559521426716118, |
|
"loss": 0.5531, |
|
"step": 15 |
|
}, |
|
{ |
|
"epoch": 1.6, |
|
"grad_norm": 0.4025287628173828, |
|
"learning_rate": 0.00019075754196709572, |
|
"loss": 0.528, |
|
"step": 20 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"grad_norm": 0.5065444707870483, |
|
"learning_rate": 0.0001842582073616649, |
|
"loss": 0.5522, |
|
"step": 25 |
|
}, |
|
{ |
|
"epoch": 2.4, |
|
"grad_norm": 0.47670507431030273, |
|
"learning_rate": 0.00017621620551276366, |
|
"loss": 0.5206, |
|
"step": 30 |
|
}, |
|
{ |
|
"epoch": 2.8, |
|
"grad_norm": 0.39348119497299194, |
|
"learning_rate": 0.00016677877587886956, |
|
"loss": 0.5305, |
|
"step": 35 |
|
}, |
|
{ |
|
"epoch": 3.2, |
|
"grad_norm": 0.5051579475402832, |
|
"learning_rate": 0.00015611870653623825, |
|
"loss": 0.5061, |
|
"step": 40 |
|
}, |
|
{ |
|
"epoch": 3.6, |
|
"grad_norm": 0.43032407760620117, |
|
"learning_rate": 0.00014443117063539038, |
|
"loss": 0.4894, |
|
"step": 45 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"grad_norm": 0.42734888195991516, |
|
"learning_rate": 0.000131930153013598, |
|
"loss": 0.4779, |
|
"step": 50 |
|
}, |
|
{ |
|
"epoch": 4.4, |
|
"grad_norm": 0.3881720006465912, |
|
"learning_rate": 0.00011884453238783185, |
|
"loss": 0.4904, |
|
"step": 55 |
|
}, |
|
{ |
|
"epoch": 4.8, |
|
"grad_norm": 0.4254741072654724, |
|
"learning_rate": 0.00010541389085854176, |
|
"loss": 0.4678, |
|
"step": 60 |
|
}, |
|
{ |
|
"epoch": 5.2, |
|
"grad_norm": 0.3979746997356415, |
|
"learning_rate": 9.18841274472569e-05, |
|
"loss": 0.4605, |
|
"step": 65 |
|
}, |
|
{ |
|
"epoch": 5.6, |
|
"grad_norm": 0.5344672799110413, |
|
"learning_rate": 7.85029559788976e-05, |
|
"loss": 0.4667, |
|
"step": 70 |
|
}, |
|
{ |
|
"epoch": 6.0, |
|
"grad_norm": 0.44632646441459656, |
|
"learning_rate": 6.551536973720298e-05, |
|
"loss": 0.4588, |
|
"step": 75 |
|
}, |
|
{ |
|
"epoch": 6.4, |
|
"grad_norm": 0.45458871126174927, |
|
"learning_rate": 5.3159155930021e-05, |
|
"loss": 0.4698, |
|
"step": 80 |
|
}, |
|
{ |
|
"epoch": 6.8, |
|
"grad_norm": 0.46564164757728577, |
|
"learning_rate": 4.16605420892506e-05, |
|
"loss": 0.4429, |
|
"step": 85 |
|
}, |
|
{ |
|
"epoch": 7.2, |
|
"grad_norm": 0.4069627523422241, |
|
"learning_rate": 3.123005411465766e-05, |
|
"loss": 0.451, |
|
"step": 90 |
|
}, |
|
{ |
|
"epoch": 7.6, |
|
"grad_norm": 0.3835998475551605, |
|
"learning_rate": 2.205866179584084e-05, |
|
"loss": 0.4238, |
|
"step": 95 |
|
}, |
|
{ |
|
"epoch": 8.0, |
|
"grad_norm": 0.4063926935195923, |
|
"learning_rate": 1.4314282383241096e-05, |
|
"loss": 0.4493, |
|
"step": 100 |
|
}, |
|
{ |
|
"epoch": 8.4, |
|
"grad_norm": 0.3951762318611145, |
|
"learning_rate": 8.138706223637827e-06, |
|
"loss": 0.4394, |
|
"step": 105 |
|
}, |
|
{ |
|
"epoch": 8.8, |
|
"grad_norm": 0.408200204372406, |
|
"learning_rate": 3.6450007480777093e-06, |
|
"loss": 0.4367, |
|
"step": 110 |
|
}, |
|
{ |
|
"epoch": 9.2, |
|
"grad_norm": 0.5846555233001709, |
|
"learning_rate": 9.154403421193225e-07, |
|
"loss": 0.429, |
|
"step": 115 |
|
}, |
|
{ |
|
"epoch": 9.6, |
|
"grad_norm": 0.3439011573791504, |
|
"learning_rate": 0.0, |
|
"loss": 0.4372, |
|
"step": 120 |
|
}, |
|
{ |
|
"epoch": 9.6, |
|
"step": 120, |
|
"total_flos": 77528353996800.0, |
|
"train_loss": 0.4846629718939463, |
|
"train_runtime": 54.9394, |
|
"train_samples_per_second": 18.202, |
|
"train_steps_per_second": 2.184 |
|
} |
|
], |
|
"logging_steps": 5, |
|
"max_steps": 120, |
|
"num_input_tokens_seen": 0, |
|
"num_train_epochs": 10, |
|
"save_steps": 500, |
|
"total_flos": 77528353996800.0, |
|
"train_batch_size": 4, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|